Beverly Wu : the official website

EEG (Electroencephalography) Workshop @ #SFU Language and Brain Lab

21 May 2011 Leave a Comment

by beverly in Linguistics, Simon Fraser University Tags: 2011, BESA, Brain, Cognitive Science, EEG, ERP, Language, Linguistics, MMN, SFU, Simon Fraser University

This weekend, Dr. Yang Zhang from the University of Minnesota visited the Language and Brain Lab at Simon Fraser University to conduct an EEG workshop. Dr. Zhang and Dr. Yue Wang (SFU LABlab) have collaborated on several projects in the past (see http://www.sfu.ca/~yuew).

*** NOTE: I joined this lab two weeks ago with only one class in undergrad neurolinguistics so when I’m trying to describe things, I don’t really actually know what I’m talking about yet. Just join me in looking at the pretty pictures and saying “OOOOOHHH that looks cool.”…especially if you actually know your cog sci. ***

So for the data acquisition and analysis part of the workshop today, I volunteered to be the test subject for the experiment. They prepped me with the EEG head net (129 electrode system):

In the EEG head net.

and then they did an auditory passive oddball paradigm / mismatch negativity sort of thing, testing passive perception of Mandarin tones and also ‘hmm’. They forgot to bring DVDs of movies with subtitles, so we put on some streaming internet TV – namely The Nature of Things, with David Suzuki.

Things I learned from being a subject for EEG experiment:

1.) My eyeballs get stressed out easily. A normal rate/style of blinking is fine – it messes up the data collected, but eye blinks are mostly easily removed in data analysis. HOWEVER, it is important to not blink too much (10 times/second) or try to hold back blinking (makes your muscles and brain signals all stressed out, and weird interference shows up) or try to blink subtly (very difficult to remove in post because the software thinks it’s a brain wave and is reluctant to remove it). Also, eye strain showed up as this weird drift when we got the mismatch negativity (MMN) chart with all the electrodes listed. Some of the little graphs representing electrodes near the front of my head would slope down to the right, and then you have to do a further pre-stimulus baseline correction on those ones. Oh well. I decided to go with contact lenses because my glasses are metal and I didn’t want them to get in the way of the head net.

2.) I really appreciate people who move their lips when they speak. David Suzuki’s The Nature of Things didn’t have subtitles, so I found myself trying to read people’s lips during the boring conversation bits. I found that if I imagined that they were actually producing sound, I could almost tell what they were saying. However, I wonder if this interfered with any of the data that was being collected. Also, I had a really stressful time trying to hold back the giggles when the show suddenly cut to a commercial of some animal doing some silly thing (I think it might’ve been one of those Telus ads), and then Ellen DeGeneres jumping around making funny faces advertising some anti-aging makeup. Then it got really gory at a few points during the show – in the Arctic section, they shot a seal, then cooked it and ate it. Then another part of the show was about commercial fishing, and then it got gory again where they gutted the fish. Oy. I get super grossed out and queasy at gory scenes. I usually only end up watching half of each House episode for this reason. Apparently the disturbances to the data being collected looked very drastic, as if I had sneezed or something.

3.) My back hurt in the plastic chair. Nothing against the chair itself, but sitting without moving for half an hour in a plastic lawn chair kinda hurts.

4.) Electrodes are goddamn picky. Before the experiment began, they had to get most of the electrodes on the head net to get a decent contact (low impedence) with my scalp. They have a map that indicates each electrode’s status – green means good, red means bad. They never went all green, no matter how much hair moving they did and electrolyte solution they dripped on the electrodes, etc. In fact, I don’t think any more than about 80% of the electrodes went green at any point during the preparation phase. Hmm.

Things I learned from EEG data analysis:

1.) Electrodes are goddamn picky. We ended up having to interpolate several electrodes from their six surrounding neighbours, because the data just looked completely wacky – like +/- way more than than the allowable 100 microvolt variance of a normal signal.

2.) Data analysis produces cool looking pictures. We used software called BESA (Brain Electrical Source Analysis) to filter and analyse the data we collected. After a whole bunch of fiddling around, we got these cool looking pictures:

MMN map for all the electrodes, created using BESA (Click to enlarge)

and then there’s a really neat topographical map function too that shows you where the positive and negative currents are at any point during your window of interest. This one is a screenshot at 137ms post-stimulus:

Topographical representation of my EEG data at 137 ms post-stimulus

3.) Data got noisier as the session progressed. You should’ve seen the difference in clean trials vs. borderline acceptable vs. rejected trials as the session progressed. I don’t know what happened – was I getting bored, tired, annoyed by the chair and fish gutting on the tv show? Was I ignoring the auditory stimuli? Did my obsession with lip reading what the TV show personalities were saying throw a wrench into it all?  These are all questions I don’t know how to answer…yet. 

4.) I’ve already forgotten how to do the data analysis we learned today. That’s okay. I’m sure I’ll learn it again soon, in detail, in the future…hopefully sometime this year, when another EEG project gets up and running in the lab.